基于阴影集的粗糙模糊可能性C均值聚类算法

相对于硬聚类算法,软聚类算法可以更好地表示具有不精确边界的类簇。粗糙集和模糊集均是用于描述不确定数据的有效的数学工具,二者互为补充。研究人员已经将粗糙集和模糊集的概念相结合,并应用到聚类算法中,提出了粗糙模糊可能性C均值聚类算法。而文中通过引入阴影集,有效地解决了粗糙模糊可能性C均值聚类算法中的阈值选择问题。

Shadowed Sets-based Rough Fuzzy Possibilistic C-means Clustering

It has been shown that soft clustering is advantageous over hard clustering in describing clusters without crisp boundaries. Both rough sets and fuzzy sets arc effective mathematical tools in handling uncertainty. As claimed in many studies,they are complementary. The theories of rough sets and fuzzy sets have been integrated into clustering algorithms,such as rough fuzzy possibilistic Gmeans clustering (RFPCM). In this study, we introduced the shadowed sets optimization theory and proposed an objective method to select the threshold in RFPCM.